Using the Molassembler python API to generate an ensemble of TS guesses from a template.
This repository contains an exemplary script (substitute.py) demonstrating how Molassembler can be used to generate ensembles of TS structures from a template.
The provided snippet requires several SCINE modules to run. SCINE Molassembler and SCINE Utilities are a core part; SCINE Database dependencies also exist, but are unused in the current version because the geometry refinement and stationary point optimization can be performed separately from the generated 3D ensembles.
The script is pointed towards templates via the templatepath variable, and writes results to the newpath variable path. It systematically replaces the bare Cp ligand in the template with the (Cp-containing) molecule provided in func_smiles. It then attempts to generate nconfs conformers (50 by default) and saves them in newpath. Both xyz files and Gaussian input files (.com) are generated; the latter can be run using Gaussian, xtb and a wrapper connecting both, such as xtb-gaussian. We note that the settings in the Gaussian input files may be optimized for the case under consideration.
This is intended to serve as an example that can be adapted to use the Molassembler library for other (similar) purposes. An example of such adaptation is given by the phosphine_conformer_generator.py, which is significantly simpler and generates conformer samples for exemplary Buchwald phosphine Pd complexes.